]>
Commit | Line | Data |
---|---|---|
3e7ee490 HJ |
1 | /* |
2 | * | |
3 | * Copyright (c) 2009, Microsoft Corporation. | |
4 | * | |
5 | * This program is free software; you can redistribute it and/or modify it | |
6 | * under the terms and conditions of the GNU General Public License, | |
7 | * version 2, as published by the Free Software Foundation. | |
8 | * | |
9 | * This program is distributed in the hope it will be useful, but WITHOUT | |
10 | * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
11 | * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for | |
12 | * more details. | |
13 | * | |
14 | * You should have received a copy of the GNU General Public License along with | |
15 | * this program; if not, write to the Free Software Foundation, Inc., 59 Temple | |
16 | * Place - Suite 330, Boston, MA 02111-1307 USA. | |
17 | * | |
18 | * Authors: | |
19 | * Haiyang Zhang <haiyangz@microsoft.com> | |
20 | * Hank Janssen <hjanssen@microsoft.com> | |
b2a5a585 | 21 | * K. Y. Srinivasan <kys@microsoft.com> |
3e7ee490 HJ |
22 | * |
23 | */ | |
0a46618d | 24 | #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt |
3e7ee490 | 25 | |
a0086dc5 GKH |
26 | #include <linux/kernel.h> |
27 | #include <linux/mm.h> | |
46a97191 | 28 | #include <linux/hyperv.h> |
011a7c3c | 29 | #include <linux/uio.h> |
9988ce68 VK |
30 | #include <linux/vmalloc.h> |
31 | #include <linux/slab.h> | |
3f335ea2 | 32 | |
0f2a6619 | 33 | #include "hyperv_vmbus.h" |
3e7ee490 | 34 | |
6fdf3b21 S |
35 | void hv_begin_read(struct hv_ring_buffer_info *rbi) |
36 | { | |
37 | rbi->ring_buffer->interrupt_mask = 1; | |
dcd0eeca | 38 | virt_mb(); |
6fdf3b21 S |
39 | } |
40 | ||
41 | u32 hv_end_read(struct hv_ring_buffer_info *rbi) | |
42 | { | |
6fdf3b21 S |
43 | |
44 | rbi->ring_buffer->interrupt_mask = 0; | |
dcd0eeca | 45 | virt_mb(); |
6fdf3b21 S |
46 | |
47 | /* | |
48 | * Now check to see if the ring buffer is still empty. | |
49 | * If it is not, we raced and we need to process new | |
50 | * incoming messages. | |
51 | */ | |
a6341f00 | 52 | return hv_get_bytes_to_read(rbi); |
6fdf3b21 S |
53 | } |
54 | ||
98fa8cf4 S |
55 | /* |
56 | * When we write to the ring buffer, check if the host needs to | |
57 | * be signaled. Here is the details of this protocol: | |
58 | * | |
59 | * 1. The host guarantees that while it is draining the | |
60 | * ring buffer, it will set the interrupt_mask to | |
61 | * indicate it does not need to be interrupted when | |
62 | * new data is placed. | |
63 | * | |
64 | * 2. The host guarantees that it will completely drain | |
65 | * the ring buffer before exiting the read loop. Further, | |
66 | * once the ring buffer is empty, it will clear the | |
67 | * interrupt_mask and re-check to see if new data has | |
68 | * arrived. | |
69 | */ | |
70 | ||
ccef9bcc S |
71 | static bool hv_need_to_signal(u32 old_write, struct hv_ring_buffer_info *rbi, |
72 | enum hv_signal_policy policy) | |
98fa8cf4 | 73 | { |
dcd0eeca | 74 | virt_mb(); |
d45faaee | 75 | if (READ_ONCE(rbi->ring_buffer->interrupt_mask)) |
98fa8cf4 S |
76 | return false; |
77 | ||
ccef9bcc S |
78 | /* |
79 | * When the client wants to control signaling, | |
80 | * we only honour the host interrupt mask. | |
81 | */ | |
82 | if (policy == HV_SIGNAL_POLICY_EXPLICIT) | |
83 | return true; | |
84 | ||
e91e84fa | 85 | /* check interrupt_mask before read_index */ |
dcd0eeca | 86 | virt_rmb(); |
98fa8cf4 S |
87 | /* |
88 | * This is the only case we need to signal when the | |
89 | * ring transitions from being empty to non-empty. | |
90 | */ | |
d45faaee | 91 | if (old_write == READ_ONCE(rbi->ring_buffer->read_index)) |
98fa8cf4 S |
92 | return true; |
93 | ||
94 | return false; | |
95 | } | |
96 | ||
822f18d4 | 97 | /* Get the next write location for the specified ring buffer. */ |
4d643114 | 98 | static inline u32 |
2b8a912e | 99 | hv_get_next_write_location(struct hv_ring_buffer_info *ring_info) |
3e7ee490 | 100 | { |
fc8c72eb | 101 | u32 next = ring_info->ring_buffer->write_index; |
3e7ee490 | 102 | |
3e7ee490 HJ |
103 | return next; |
104 | } | |
105 | ||
822f18d4 | 106 | /* Set the next write location for the specified ring buffer. */ |
3e7ee490 | 107 | static inline void |
2b8a912e | 108 | hv_set_next_write_location(struct hv_ring_buffer_info *ring_info, |
fc8c72eb | 109 | u32 next_write_location) |
3e7ee490 | 110 | { |
fc8c72eb | 111 | ring_info->ring_buffer->write_index = next_write_location; |
3e7ee490 HJ |
112 | } |
113 | ||
822f18d4 | 114 | /* Get the next read location for the specified ring buffer. */ |
4d643114 | 115 | static inline u32 |
2b8a912e | 116 | hv_get_next_read_location(struct hv_ring_buffer_info *ring_info) |
3e7ee490 | 117 | { |
fc8c72eb | 118 | u32 next = ring_info->ring_buffer->read_index; |
3e7ee490 | 119 | |
3e7ee490 HJ |
120 | return next; |
121 | } | |
122 | ||
b2a5a585 | 123 | /* |
b2a5a585 | 124 | * Get the next read location + offset for the specified ring buffer. |
822f18d4 | 125 | * This allows the caller to skip. |
b2a5a585 | 126 | */ |
4d643114 | 127 | static inline u32 |
2b8a912e | 128 | hv_get_next_readlocation_withoffset(struct hv_ring_buffer_info *ring_info, |
1ac58644 | 129 | u32 offset) |
3e7ee490 | 130 | { |
fc8c72eb | 131 | u32 next = ring_info->ring_buffer->read_index; |
3e7ee490 | 132 | |
fc8c72eb HZ |
133 | next += offset; |
134 | next %= ring_info->ring_datasize; | |
3e7ee490 HJ |
135 | |
136 | return next; | |
137 | } | |
138 | ||
822f18d4 | 139 | /* Set the next read location for the specified ring buffer. */ |
3e7ee490 | 140 | static inline void |
2b8a912e | 141 | hv_set_next_read_location(struct hv_ring_buffer_info *ring_info, |
fc8c72eb | 142 | u32 next_read_location) |
3e7ee490 | 143 | { |
fc8c72eb | 144 | ring_info->ring_buffer->read_index = next_read_location; |
ab028db4 | 145 | ring_info->priv_read_index = next_read_location; |
3e7ee490 HJ |
146 | } |
147 | ||
822f18d4 | 148 | /* Get the size of the ring buffer. */ |
4d643114 | 149 | static inline u32 |
2b8a912e | 150 | hv_get_ring_buffersize(struct hv_ring_buffer_info *ring_info) |
3e7ee490 | 151 | { |
fc8c72eb | 152 | return ring_info->ring_datasize; |
3e7ee490 HJ |
153 | } |
154 | ||
822f18d4 | 155 | /* Get the read and write indices as u64 of the specified ring buffer. */ |
59471438 | 156 | static inline u64 |
2b8a912e | 157 | hv_get_ring_bufferindices(struct hv_ring_buffer_info *ring_info) |
3e7ee490 | 158 | { |
fc8c72eb | 159 | return (u64)ring_info->ring_buffer->write_index << 32; |
3e7ee490 HJ |
160 | } |
161 | ||
8f1136ae | 162 | /* |
8f1136ae S |
163 | * Helper routine to copy to source from ring buffer. |
164 | * Assume there is enough room. Handles wrap-around in src case only!! | |
8f1136ae S |
165 | */ |
166 | static u32 hv_copyfrom_ringbuffer( | |
167 | struct hv_ring_buffer_info *ring_info, | |
168 | void *dest, | |
169 | u32 destlen, | |
170 | u32 start_read_offset) | |
171 | { | |
172 | void *ring_buffer = hv_get_ring_buffer(ring_info); | |
173 | u32 ring_buffer_size = hv_get_ring_buffersize(ring_info); | |
174 | ||
175 | u32 frag_len; | |
176 | ||
177 | /* wrap-around detected at the src */ | |
178 | if (destlen > ring_buffer_size - start_read_offset) { | |
179 | frag_len = ring_buffer_size - start_read_offset; | |
180 | ||
181 | memcpy(dest, ring_buffer + start_read_offset, frag_len); | |
182 | memcpy(dest + frag_len, ring_buffer, destlen - frag_len); | |
183 | } else | |
184 | ||
185 | memcpy(dest, ring_buffer + start_read_offset, destlen); | |
186 | ||
187 | ||
188 | start_read_offset += destlen; | |
189 | start_read_offset %= ring_buffer_size; | |
190 | ||
191 | return start_read_offset; | |
192 | } | |
193 | ||
194 | ||
7581578d | 195 | /* |
7581578d S |
196 | * Helper routine to copy from source to ring buffer. |
197 | * Assume there is enough room. Handles wrap-around in dest case only!! | |
7581578d S |
198 | */ |
199 | static u32 hv_copyto_ringbuffer( | |
fc8c72eb HZ |
200 | struct hv_ring_buffer_info *ring_info, |
201 | u32 start_write_offset, | |
202 | void *src, | |
7581578d S |
203 | u32 srclen) |
204 | { | |
205 | void *ring_buffer = hv_get_ring_buffer(ring_info); | |
206 | u32 ring_buffer_size = hv_get_ring_buffersize(ring_info); | |
207 | u32 frag_len; | |
208 | ||
209 | /* wrap-around detected! */ | |
210 | if (srclen > ring_buffer_size - start_write_offset) { | |
211 | frag_len = ring_buffer_size - start_write_offset; | |
212 | memcpy(ring_buffer + start_write_offset, src, frag_len); | |
213 | memcpy(ring_buffer, src + frag_len, srclen - frag_len); | |
214 | } else | |
215 | memcpy(ring_buffer + start_write_offset, src, srclen); | |
3e7ee490 | 216 | |
7581578d S |
217 | start_write_offset += srclen; |
218 | start_write_offset %= ring_buffer_size; | |
219 | ||
220 | return start_write_offset; | |
221 | } | |
3e7ee490 | 222 | |
822f18d4 | 223 | /* Get various debug metrics for the specified ring buffer. */ |
a75b61d5 | 224 | void hv_ringbuffer_get_debuginfo(struct hv_ring_buffer_info *ring_info, |
80682b7a | 225 | struct hv_ring_buffer_debug_info *debug_info) |
3e7ee490 | 226 | { |
fc8c72eb HZ |
227 | u32 bytes_avail_towrite; |
228 | u32 bytes_avail_toread; | |
3e7ee490 | 229 | |
fc8c72eb | 230 | if (ring_info->ring_buffer) { |
2b8a912e | 231 | hv_get_ringbuffer_availbytes(ring_info, |
fc8c72eb HZ |
232 | &bytes_avail_toread, |
233 | &bytes_avail_towrite); | |
3e7ee490 | 234 | |
fc8c72eb HZ |
235 | debug_info->bytes_avail_toread = bytes_avail_toread; |
236 | debug_info->bytes_avail_towrite = bytes_avail_towrite; | |
82f8bd40 | 237 | debug_info->current_read_index = |
fc8c72eb | 238 | ring_info->ring_buffer->read_index; |
82f8bd40 | 239 | debug_info->current_write_index = |
fc8c72eb | 240 | ring_info->ring_buffer->write_index; |
82f8bd40 | 241 | debug_info->current_interrupt_mask = |
fc8c72eb | 242 | ring_info->ring_buffer->interrupt_mask; |
3e7ee490 HJ |
243 | } |
244 | } | |
245 | ||
822f18d4 | 246 | /* Initialize the ring buffer. */ |
72a95cbc | 247 | int hv_ringbuffer_init(struct hv_ring_buffer_info *ring_info, |
9988ce68 | 248 | struct page *pages, u32 page_cnt) |
3e7ee490 | 249 | { |
9988ce68 VK |
250 | int i; |
251 | struct page **pages_wraparound; | |
252 | ||
253 | BUILD_BUG_ON((sizeof(struct hv_ring_buffer) != PAGE_SIZE)); | |
3e7ee490 | 254 | |
fc8c72eb | 255 | memset(ring_info, 0, sizeof(struct hv_ring_buffer_info)); |
3e7ee490 | 256 | |
9988ce68 VK |
257 | /* |
258 | * First page holds struct hv_ring_buffer, do wraparound mapping for | |
259 | * the rest. | |
260 | */ | |
261 | pages_wraparound = kzalloc(sizeof(struct page *) * (page_cnt * 2 - 1), | |
262 | GFP_KERNEL); | |
263 | if (!pages_wraparound) | |
264 | return -ENOMEM; | |
265 | ||
266 | pages_wraparound[0] = pages; | |
267 | for (i = 0; i < 2 * (page_cnt - 1); i++) | |
268 | pages_wraparound[i + 1] = &pages[i % (page_cnt - 1) + 1]; | |
269 | ||
270 | ring_info->ring_buffer = (struct hv_ring_buffer *) | |
271 | vmap(pages_wraparound, page_cnt * 2 - 1, VM_MAP, PAGE_KERNEL); | |
272 | ||
273 | kfree(pages_wraparound); | |
274 | ||
275 | ||
276 | if (!ring_info->ring_buffer) | |
277 | return -ENOMEM; | |
278 | ||
fc8c72eb HZ |
279 | ring_info->ring_buffer->read_index = |
280 | ring_info->ring_buffer->write_index = 0; | |
3e7ee490 | 281 | |
822f18d4 | 282 | /* Set the feature bit for enabling flow control. */ |
046c7911 S |
283 | ring_info->ring_buffer->feature_bits.value = 1; |
284 | ||
9988ce68 VK |
285 | ring_info->ring_size = page_cnt << PAGE_SHIFT; |
286 | ring_info->ring_datasize = ring_info->ring_size - | |
287 | sizeof(struct hv_ring_buffer); | |
3e7ee490 | 288 | |
fc8c72eb | 289 | spin_lock_init(&ring_info->ring_lock); |
3e7ee490 HJ |
290 | |
291 | return 0; | |
292 | } | |
293 | ||
822f18d4 | 294 | /* Cleanup the ring buffer. */ |
2dba688b | 295 | void hv_ringbuffer_cleanup(struct hv_ring_buffer_info *ring_info) |
3e7ee490 | 296 | { |
9988ce68 | 297 | vunmap(ring_info->ring_buffer); |
3e7ee490 HJ |
298 | } |
299 | ||
822f18d4 | 300 | /* Write to the ring buffer. */ |
633c4dce | 301 | int hv_ringbuffer_write(struct hv_ring_buffer_info *outring_info, |
ccef9bcc S |
302 | struct kvec *kv_list, u32 kv_count, bool *signal, bool lock, |
303 | enum hv_signal_policy policy) | |
3e7ee490 | 304 | { |
4408f531 | 305 | int i = 0; |
fc8c72eb | 306 | u32 bytes_avail_towrite; |
fc8c72eb | 307 | u32 totalbytes_towrite = 0; |
3e7ee490 | 308 | |
66a60543 | 309 | u32 next_write_location; |
98fa8cf4 | 310 | u32 old_write; |
fc8c72eb | 311 | u64 prev_indices = 0; |
fe760e4d | 312 | unsigned long flags = 0; |
3e7ee490 | 313 | |
011a7c3c S |
314 | for (i = 0; i < kv_count; i++) |
315 | totalbytes_towrite += kv_list[i].iov_len; | |
3e7ee490 | 316 | |
fc8c72eb | 317 | totalbytes_towrite += sizeof(u64); |
3e7ee490 | 318 | |
fe760e4d S |
319 | if (lock) |
320 | spin_lock_irqsave(&outring_info->ring_lock, flags); | |
3e7ee490 | 321 | |
a6341f00 | 322 | bytes_avail_towrite = hv_get_bytes_to_write(outring_info); |
3e7ee490 | 323 | |
822f18d4 VK |
324 | /* |
325 | * If there is only room for the packet, assume it is full. | |
326 | * Otherwise, the next time around, we think the ring buffer | |
327 | * is empty since the read index == write index. | |
328 | */ | |
fc8c72eb | 329 | if (bytes_avail_towrite <= totalbytes_towrite) { |
fe760e4d S |
330 | if (lock) |
331 | spin_unlock_irqrestore(&outring_info->ring_lock, flags); | |
d2598f01 | 332 | return -EAGAIN; |
3e7ee490 HJ |
333 | } |
334 | ||
454f18a9 | 335 | /* Write to the ring buffer */ |
2b8a912e | 336 | next_write_location = hv_get_next_write_location(outring_info); |
3e7ee490 | 337 | |
98fa8cf4 S |
338 | old_write = next_write_location; |
339 | ||
011a7c3c | 340 | for (i = 0; i < kv_count; i++) { |
2b8a912e | 341 | next_write_location = hv_copyto_ringbuffer(outring_info, |
fc8c72eb | 342 | next_write_location, |
011a7c3c S |
343 | kv_list[i].iov_base, |
344 | kv_list[i].iov_len); | |
3e7ee490 HJ |
345 | } |
346 | ||
454f18a9 | 347 | /* Set previous packet start */ |
2b8a912e | 348 | prev_indices = hv_get_ring_bufferindices(outring_info); |
3e7ee490 | 349 | |
2b8a912e | 350 | next_write_location = hv_copyto_ringbuffer(outring_info, |
fc8c72eb HZ |
351 | next_write_location, |
352 | &prev_indices, | |
b219b3f7 | 353 | sizeof(u64)); |
3e7ee490 | 354 | |
98fa8cf4 | 355 | /* Issue a full memory barrier before updating the write index */ |
dcd0eeca | 356 | virt_mb(); |
3e7ee490 | 357 | |
454f18a9 | 358 | /* Now, update the write location */ |
2b8a912e | 359 | hv_set_next_write_location(outring_info, next_write_location); |
3e7ee490 | 360 | |
3e7ee490 | 361 | |
fe760e4d S |
362 | if (lock) |
363 | spin_unlock_irqrestore(&outring_info->ring_lock, flags); | |
98fa8cf4 | 364 | |
ccef9bcc | 365 | *signal = hv_need_to_signal(old_write, outring_info, policy); |
3e7ee490 HJ |
366 | return 0; |
367 | } | |
368 | ||
940b68e2 VK |
369 | int hv_ringbuffer_read(struct hv_ring_buffer_info *inring_info, |
370 | void *buffer, u32 buflen, u32 *buffer_actual_len, | |
371 | u64 *requestid, bool *signal, bool raw) | |
3e7ee490 | 372 | { |
fc8c72eb HZ |
373 | u32 bytes_avail_toread; |
374 | u32 next_read_location = 0; | |
375 | u64 prev_indices = 0; | |
940b68e2 VK |
376 | struct vmpacket_descriptor desc; |
377 | u32 offset; | |
378 | u32 packetlen; | |
379 | int ret = 0; | |
3e7ee490 | 380 | |
fc8c72eb | 381 | if (buflen <= 0) |
a16e1485 | 382 | return -EINVAL; |
3e7ee490 | 383 | |
3e7ee490 | 384 | |
940b68e2 VK |
385 | *buffer_actual_len = 0; |
386 | *requestid = 0; | |
387 | ||
a6341f00 | 388 | bytes_avail_toread = hv_get_bytes_to_read(inring_info); |
454f18a9 | 389 | /* Make sure there is something to read */ |
940b68e2 VK |
390 | if (bytes_avail_toread < sizeof(desc)) { |
391 | /* | |
392 | * No error is set when there is even no header, drivers are | |
393 | * supposed to analyze buffer_actual_len. | |
394 | */ | |
3eba9a77 | 395 | return ret; |
940b68e2 | 396 | } |
3e7ee490 | 397 | |
940b68e2 VK |
398 | next_read_location = hv_get_next_read_location(inring_info); |
399 | next_read_location = hv_copyfrom_ringbuffer(inring_info, &desc, | |
400 | sizeof(desc), | |
401 | next_read_location); | |
402 | ||
403 | offset = raw ? 0 : (desc.offset8 << 3); | |
404 | packetlen = (desc.len8 << 3) - offset; | |
405 | *buffer_actual_len = packetlen; | |
406 | *requestid = desc.trans_id; | |
407 | ||
3eba9a77 S |
408 | if (bytes_avail_toread < packetlen + offset) |
409 | return -EAGAIN; | |
940b68e2 | 410 | |
3eba9a77 S |
411 | if (packetlen > buflen) |
412 | return -ENOBUFS; | |
3e7ee490 | 413 | |
1ac58644 | 414 | next_read_location = |
2b8a912e | 415 | hv_get_next_readlocation_withoffset(inring_info, offset); |
3e7ee490 | 416 | |
2b8a912e | 417 | next_read_location = hv_copyfrom_ringbuffer(inring_info, |
fc8c72eb | 418 | buffer, |
940b68e2 | 419 | packetlen, |
fc8c72eb | 420 | next_read_location); |
3e7ee490 | 421 | |
2b8a912e | 422 | next_read_location = hv_copyfrom_ringbuffer(inring_info, |
fc8c72eb | 423 | &prev_indices, |
4408f531 | 424 | sizeof(u64), |
fc8c72eb | 425 | next_read_location); |
3e7ee490 | 426 | |
822f18d4 VK |
427 | /* |
428 | * Make sure all reads are done before we update the read index since | |
429 | * the writer may start writing to the read area once the read index | |
430 | * is updated. | |
431 | */ | |
dcd0eeca | 432 | virt_mb(); |
3e7ee490 | 433 | |
454f18a9 | 434 | /* Update the read index */ |
2b8a912e | 435 | hv_set_next_read_location(inring_info, next_read_location); |
3e7ee490 | 436 | |
a389fcfd | 437 | *signal = hv_need_to_signal_on_read(inring_info); |
c2b8e520 | 438 | |
940b68e2 | 439 | return ret; |
b5f53dde | 440 | } |